Lamellar Mesophase of Poly(ethy1ene oxide) - American Chemical

Aug 9, 1994 - Department of Solid State Physics, Rise National Laboratory, DK-4000 Roskilde, Denmark. W y n Brown. Department of Physical Chemistry, ...
0 downloads 0 Views 1MB Size
Macromolecules 1996,28, 1458- 1463

1458

Lamellar Mesophase of Poly(ethy1ene oxide)-Poly(propy1ene oxide)-Poly(ethy1ene oxide) Melts and Water-Swollen Mixtures Kell Mortensen* Department of Solid State Physics, Rise National Laboratory, DK-4000 Roskilde, Denmark

W y n Brown Department of Physical Chemistry, Uppsala University, S-75121Uppsala, Sweden

Erling Jgrgensen Chemistry Department, Roskilde University Center, Roskilde, Denmark Received August 9,1994; Revised Manuscript Received November 21, 1994@ ABSTRACT: Triblock copolymers of poly(ethy1ene oxide) (PEO) and poly(propy1ene oxide) (PPO) form lamellar mesophases at low temperatures. The order-to-disorder transition is closely related to the melting transition of the PEO subunit. When water is incorporated into the copolymer melt, a single phase is formed at high temperatures, which is possibly a lamellar mesophase, formed as a consequence of the hydrophobic PPO blocks. At low temperatures two phases are present, one of which is similar to the high temperature phase while the other is similar to the lamellar melt, but swollen up to 10%.

I. Introduction The phase behavior of block copolymers has recently attracted great interest, as a result of both the commercial utility and the novel physical properties. Especially, many studies have been performed on amorphous diblock copolymers1 in which the mesophase formation is determined by the Flory-Huggins interaction parameter, fluctuations, and conformational symmetry. Mesophase formation can also be driven by other effects, as, for example, the tendency of one of the blocks to crystallize, which will be shown below. Studies of systems of crystalline blocks are rather rare and not well understood. Such polymers have, however, great technological importance. A number of commercial block copolymers are based on poly(ethy1ene oxide) which crystallizes at a low temperature depending on the molecular weight as well as the structure of the surrounding blocks. Such copolymers are, for example, utilized as polymer surfactants for viscosity control, etc. It is therefore important to know how the incorporation of a solvent affects the crystalline mesophase of the melt. In the present study, we describe an experimental study of triblock copolymers of poly(ethy1ene oxide) and poly(propy1ene oxide), both in the melt and with water as a diluent. The mesophase of the melt appears to be lamellar. Upon increasing the water content, the mesophase melting (order-to-disorder) transition temperature decreases. Moreover, the lamellar mesophase is swollen to some limiting value, when the solvent is introduced. 11. Experimental Section A. Material. The triblock copolymer, polytethylene oxidelpoly(propy1ene oxidel-poly(ethy1ene oxide),

or PEO-PPO-PEO, abbreviated P85, was obtained from BASF Corp., Wyandotte, MI, and used without further purification. The material has a declared molecular weight of @Abstractpublished in Advance ACS Abstracts, February 1, 1995.

4500,2200 for the PEO component and 2300 for PPO. The solutions were prepared at ambient temperature and then heated to roughly 60 "C, to make homogeneous samples. Deuterium oxide, DzO, was used in order to get good contrast and a low background in the neutron scattering experiments. The solutions discussed below are all given in weight percent (wt %).

B. Small-AngleNeutron Scattering. Small-angle neutron scattering experiments were performed using the RiseSANS facility, which is a flexible instrument covering scatwith variable neutron tering vectors from 0.002 to 0.5 kl, wavelength resolution. The samples were mounted in sealed quartz containers (Suprasil from Hellma, FRG), with a 1mm flight path. The experiments were erformed using neutron wavelengths (1)of 3, 6, and 10 , with 1, 3, and 6 m sample-todetector distances, respectively. The neutron wavelength resolution was AU1 = 0.18, the neutron beam collimation was determined by the pinhole sizes of 16 and 7 mm diameter at the source and sample positions, respectively, and collimation lengths were equal t o the sample-to-detector distance. The scattering data were corrected for the background arising from the quartz cell and from other sources, as measured with the neutron beam blocked by plastic containing boron at the sample position. The incoherent scattering from HzO was used to determine deviations from a uniform detector response and to convert the data into absolute units. The scattering patterns discussed in the present paper are all azimuthally isotropic. The data have been reduced by azimuthally averaging to the one-dimensional Z(q) scattering functions which are only dependent on the absolute value of tj, where tj is given by the scattering angle 0 and the neutron wavelength 2: IqI = q = (4n/2) sin(O/2). C. Light Scattering. Dynamic light scattering measurements were made on the P85 melt as well as on concentrated P85 solutions. The scattering cells (10-mL sealed cylindrical ampules) were immersed in a large-diameter thermostated bath of index-matching liquid (silicon oil). The polarized (VV) DLS measurements, in the self-beating (homodyne) mode, were performed using a frequency-stabilized Coherent Innova Ar-ion laser operating at 488 nm with adjustable output power. The light was vertically polarized with a Glan-Thompson polarizer, with an extinction coefficient better than The detector optics employed a 4-pm-diameter monomodal fiber coupled to an ITT FW130 photomultiplier the output of which was digitized by an ALV-5000 digital multiple-.t autocorrelator, Langen GmbH, with 288 exponentially spaced channels. It has a minimum real time sampling time of 0.2 ,us and a

1

0 1995 American Chemical Society 0024-9297/95/2228-1458$09.00/0

Macromolecules, Vol. 28, No. 5, 1995

Lamellar Mesophase of PEO-PPO-PEO Melts 1459

maximum of about 100 s. The intensity autocorrelation function, g W ) , was measured at different angles. In the present study, in most cases, the temperature was 80 "Cand was controlled to within k0.02 "C. Measurements of depolarized dynamic light scattering were also made where the scattered light passed through a GlanThompson polarizer with an extinction coefficient better than lo-', whose orientation was adjusted to give the minimum intensity for a dilute solution of a high molecular weight polystyrene in ethyl acetate. The DLS data were analyzed by nonlinear regression procedures. The various models used in the fitting procedures are expressed with respect tog'lft), while the fitting was performed with respect to the measuredg"W), described as

pa5

I

q (1/N

where /3 is a nonideality factor which accounts for the deviation from ideal correlation. g")(t)can be written as the Laplace transform of the distribution of relaxation rates, G(r):

Figure 1. Neutron scattering function of the P85 melt as obtained at temperatures between T = 7 "C and T = 80 "C. P85 LORENTZ FIT

6.0

where r is the relaxation rate and t is the lag time. For relaxation times, t, eq 2 will be expressed as

>

6 Bc

z Y 2 a where tA(t) TG(T) in the logarithmic scale. tA(t)was obtained by regularized inverse Laplace transformation of the dynamic light scattering data using a constrained regularization calculation algorithm called REPES, as incorporated in the analysis package GENDIST (see ref 9). This algorithm directly minimizes the sum of the squared differences between functions. It allows the experimental and calculated d2)(t) selection of a "smoothing parameter", probability to reject (the higher the probability t o reject, the greater the smoothing). A value of 0.5 was chosen in all analyses. It is remarked that, in the present studies, there is a n insignificant difference between the correlogram shapes and the positioning of the depolarized (VH) and polarized (VV) correlograms on the log-time axis. The temperature of 80 "C was selected for the majority of the DLS measurements since preliminary studies showed that glass-clear mixtures with water existed at this temperature over the whole concentration range up to and including the melt. The light source for the static light scattering was a 3-mW He-Ne laser (1= 633 nm). The optical constant for vertically polarized light is

K = 4m,(dnldc)21NA~4 where no is the solvent refractive index, dnldc the refractive index increment which was measured using a specially constructed differential refractometer with Rayleigh interference optics ((dnldc) = 0.131 m u g at 633 nm), and N A is Avogadro's number. RO is the Rayleigh ratio obtained by calibration measurements with benzene: R w ~= 11.85 x cm-l at 25 "C. D. Viscosity Measurements. Steady shear viscosities were measured a t 60, 70, and 80 "C using a Bohlin VOR rheometer (Lund, Sweden). The Couette geometry ((214)was used at shear rates between 0.1 and 100 s-l. The measured viscosities were independent of shear rate in the investigated range.

111. Results and Discussion

k PEO-PPO-PEO Melt. 1. Neutron Scattering. Figure 1 shows the scattering function of the P85 melt as observed at temperatures between T = 7 "C and T = 80 "C. The scattering function clearly reveals lamellar structure a t low temperature, with both first- and

a

4.0

2.0

0.0 TEMPERATURE (C) P 8 5 LORENTZ FIT

0.04

0.03

2 \

b

1

1

5 0.02

Macromolecules, Vol. 28, No. 5, 1995

1460 Mortensen et al. Above T, = 40 "C, no correlation peak is observable within the experimental resolution (Figure 1). This is quite different from observations on the amorphous block copolymers, where there is typically only a relatively small change in peak intensity on crossing the order-disorder transition (ODT). Above the ODT, a pronounced Leibler peak appears due to spatial concentration f l ~ c t u a t i o n s . ~ The peak intensity is, however, given not only by the amplitude of the concentration fluctuations but also by the scattering contrast between the two types of polymer blocks. In the P85 block copolymer melt, there is only significant contrast between PEO and PPO because the PEO block is crystalline whereas PPO is amorphous, thus resulting in a marked mass-density difference. At T > T, = 40 "C,PEO is also amorphous, and the scattering contrast vanishes. It is therefore not possible to detect eventual composition fluctuations above the PEO melting temperature with small-angle neutron scattering. Light scattering, as shown below, gives, however, an indication of important composition fluctuations in the disordered phase. Future studies should include partly deuterated copolymers, in which sufficient contrast for neutron scattering experiments can also be obtained in the amorphous regime. 2 . Light Scattering. The P85 melt has a strongly anisotropic light scattering signal. Dynamic light scattering (DLS) measurements were made as a function of angle in both the Vv and VHgeometries. The VVand VH results were essentially identical, a result which is ~ temperature expected due to strong ~ o u p l i n g .The range investigated covered in 10" steps the interval from 40 to 120 "C. Figure 3a shows the VH intensityintensity correlograms for the melt at 80 'C and Figure 3b shows the corresponding Laplace inversion results. The correlograms are approximately q2-dependent(Figure 3c), as was also observed t o be the case with the "inverse" copolymer conformation, Pluronic-R Furthermore, the correlograms are single exponential (the exponent for fits t o a stretched exponential is very close to unity), whereas most polymeric melts, which exhibit anisotropy due to segmental orientation relaxation, are typified by broad, q-independent, relaxation distributions. The most ready interpretation is that the large anisotropy derives from the microstructure of domains which are rather homogeneous in overall size and which can move diffusively in the melt. A similar description was recently given by Fytas et aL6 for an asymmetric diblock copolymer in solution close to the order-todisorder transition (ODT) temperature. The temperature dependence of the relaxation rate in the melt is small and should primarily depend on the viscosity of the medium. The latter (expressed in Pa s) is given for the melt over the limited range 60-80 'C by the relationship: ln(q) = -12.175

+ 3707T1

(4)

where T is in Kelvin. Together with the macroscopic viscosity, it is possible to estimate an apparent domain dimension using the Stokes-Einstein equation. With 7 = 0.188 Pa s (at 80 "C), (Rh)app = 220 nm. This is a n estimate of the size of the "islands" moving in the melt and which are probably present in low number density. For this reason (Rh)app should be close to the true size which would exist in the absence of interparticle interactions.

-2

0

2

6

4

0

log(t/rs)

t

10 t

5

4

6

,

l

/

lop(?/ ps) I

,

,

,

7

,

C

4 l

0

0,

-rz k

4 2

I

0l 0 . 0 0 . 1 0.2 0.3 0 . 4 0.5 0.6 0.7 0 . 6L 0.9

sin' (8/2) Figure 3. (a) Depolarized correlograms for the P85 melt at 80 "C at different angles. Peaks from right to left: (1)60"; (2) 75"; (3) 90"; (4) 110";(5) 130". (b) Inverse Laplace transformation of the correlograms in Figure 3a. Peaks from right t o left: (1)60 "; (2) 75"; (3) 90"; (4)110"; (5) 130". (c) Plot of the relaxation rate (T/s-') versus sin2 (W2) obtained from the ILT analyses of the correlograms in Figure 3a.

B. PEO-PPO-PEO Melt Incorporated with Water. I. Neutron Scattering. Incorporating small amounts of water (